Large scale DNA sequence efforts have resulted in knowledge of the amino acid sequences of a large number of proteins, many of which have no known function. Since X-ray crystallography and other physical and biochemical methods for characterizing proteins are laborious and slow, there is great interest int he use of computers to help determine structural correlates of function and to ascertain relationships between known proteins and unknown ones. Recent progress has been made by application of artificial intelligence techniques resulting in a protein analysis system known as ARIADNE. We propose here to explore the feasibility of implementing ARIADNE in a form that can be used by individual laboratories and scientists on personal computers or workstations. We plan to implement ARIADNE on a personal computer, provide it with a user friendly interface and seamless access to molecular databases, and evaluate its performance. As a potential commercial product such software could be used in many in research labs. The program could also be a very important tool for data resulting from large scale sequencing efforts such as the Human Genome Project.